Topic: Train-Test Splits and Cross-Validation

Refitting a Selected Procedure Before Test Evaluation

Random Splits and Reproducible Seeds

Nested Cross-Validation for Model Selection

Leave-One-Out Cross-Validation

Imputation Within Resampling

Fitting Preprocessing Steps Inside Each Fold

Hand-Calculating Five-Fold Cross-Validation

Diagnosing Variation Across Folds

Detecting Leakage from Full-Data Preprocessing

Constructing a Holdout Evaluation